Keeping an IEP alive: a named owner, a trigger register, and why the calendar is the weakest safeguard you have

You built the Integrated Evidence Plan, and you built it well. You ran the gap analysis, ranked what actually mattered, and got clinical, medical and the early commercial folks into one room and out the other side with a plan everyone believed in. It carried you through the thing it was built for: the raise, the End-of-Phase-II meeting, the board update. (If you have not built one yet, start here is where you begin; this post is about what happens next.) And because you are conscientious, you did one more responsible thing: you set a reminder to revisit it. Quarterly, say. Box ticked.

That quarterly reminder is not the governance safeguard you think it is. A calendar is the weakest trigger you can hang a living plan on, and it is a large part of why the plan quietly stopped informing decisions two months after you delivered it.

I sat in a portfolio review a while ago where the IEP on screen still listed the lead competitor as Phase II. That competitor had read out Phase III data months earlier. Everyone in the room knew it. Nobody had told the plan, because telling the plan was nobody's job.

It was not wrong through negligence.

It was wrong because it had been finished.

That is the failure worth taking apart, and it will not yield to anyone trying harder. Nobody in that room was slacking; the plan's own architecture did the damage on its own. Two questions decide it: who owns the plan, and what makes it move.

The "delivery fallacy"

Call it the "delivery fallacy": the quiet assumption that the day you hand the plan over is the day the work is done. The IEP was scoped as a deliverable for one triggering event. Written, presented, filed. And once that event had passed, updating it sat in precisely no one's job description.

This is not a fringe complaint. Two independent industry sources name the same failure mode, for the same artefact, in almost the same breath. The Medical Affairs Professional Society's 2025 guidance on integrated evidence-generation planning holds that the value of such a plan lies in the outcomes it achieves, and warns against treating a plan as "fixed and simply requiring implementation" [1]. ZS Associates puts it more bluntly: rather than letting an IEP "sit on a shelf," it should be "a mechanism to execute, track and maintain" [2].

The plan is not the deliverable. The deliverable was only ever a snapshot of it the plan as it stood the day it shipped.

So the fix cannot be "try harder to remember." Memory is exactly what failed. The fix is structural: an owner, and a defined set of triggers, so the plan gets refreshed without anyone relying on their own good intentions.

Why is the calendar the wrong review cadence for an IEP?

A fixed review interval fails in two directions at once. Set it too far out and you miss what mattered: a comparator reads out pivotal data in March, your review is booked for November, and for eight months the plan quietly misinforms every decision that leans on it. Set it too frequent and you burn scarce attention on quarters where nothing moved, which trains the team to treat the review as a formality. And a formality is something nobody prepares for.

The field that manages decaying evidence for a living has already worked this out. Living systematic reviews and living guidelines exist to keep an evidence synthesis current as new evidence arrives, instead of freezing it and revisiting on a fixed schedule. The Australian living guidelines taskforce, which ran weekly-updated national COVID-19 guidelines through the pandemic, put it cleanly: decisions about how often new evidence should be considered "should be driven by the rate of emergence of new evidence, clinical uncertainty and importance of the area under review" [3]. Rate of change sets the cadence. Not the calendar.

An honest complication now, because this literature does not speak with one voice. Shojania and colleagues, in 2007, found a signal for updating in 57% of the systematic reviews they tracked, appearing inside a year for around 15% [4]. But a signal is not a reversal. French (2005) and Jaidee (2010) measured how often updating actually changed the conclusion, and found it far rarer: 9% and 3.7% respectively [5][6]. Those are different questions. "Something worth noticing appeared" is not "the bottom line flipped," and you should not quote one as if it were the other. That said, Shojania studied systematic reviews, an older and different object than a biotech's evidence plan, so treat the number as an analogy, never as a figure pinned to your IEP. (Living-guideline methodology has active critics among major medical societies, so this is a live debate, not settled doctrine.) The field's own resolution is the useful lesson: review triggered by priority and by signal, not a blanket high-frequency schedule and not a long fixed interval.

Here is the model that fits. An IEP is a forecast, not a photograph. It is accurate the morning it is issued and it decays as the landscape shifts beneath it. When a new front appears on the chart, a meteorologist re-runs the model; they do not wait for the anniversary of the last run to discover the storm changed course. Set your cadence by the weather, not by the date.

Fix one: put a name against the plan

Governance sounds like it should mean a committee and a charter. It does not. For a resource-tight biotech it means one thing first. A name. One person with explicit authority to declare that the plan needs updating and to convene the refresh. ZS Associates describes exactly this: "an assigned, accountable project owner who can make swift decisions" [2].

Why does naming one person matter so much? Look at how the mature living-guideline programmes are built. The WHO's living guideline on covid-19 drugs runs a standing steering committee that watches for triggers and a separate guideline development group that executes the update [7]. Trigger-owner and executor are deliberately split, and both are named. Nobody is left assuming someone else is watching.

That assumption is the whole problem, and there is a piece of social psychology that illuminates it, offered strictly as an analogy. Darley and Latané's work on the diffusion of responsibility found a person in trouble was helped around 85% of the time when someone believed they alone could act, 62% with one other person present, and only 31% in the largest group [8]. The more people who could act, the less likely any single one did. "The team owns it" is the organisational version of a crowded room. Nobody's job. So nobody does it.

Who should hold the name? Several sources point to medical affairs as the natural long-term owner. ZS's April 2022 survey of more than thirty leaders across fifteen of the top thirty pharma companies found medical affairs oversees the IEP at 67% of them, with 47% running a dedicated centre of excellence [2]. For a small biotech with no medical affairs function yet, the principle still holds: the owner is whoever stays accountable for the evidence story once the current triggering event is long behind you. This is the governance that makes the handoff from clinical development to medical affairs at launch real rather than aspirational, turning the IEP into the living strategy it is meant to be.

The Monday version: put one name against the plan, and give that person the authority to call a refresh. Without the authority, the name is decoration.

Fix two: a trigger register, with the calendar demoted to backstop

A name alone will not keep a plan current; it needs something to pull the trigger.

Write down, inside the plan itself, a short list of events that each oblige the owner to open a review. Call it a trigger register. It converts "we noticed that" into "we are obliged to act on that," which is the entire difference between a plan that stays current and one that drifts.

MAPS's 2025 guidance offers a ready-made four-category taxonomy of what to watch, and it maps neatly onto what a biotech already tracks [1]:

• Competitive landscape: a comparator's pivotal readout, approval, label change or programme failure. Any one of these can reprice your differentiation overnight.
• Standards and guidelines: a shift in treatment guidelines or in standard of care that moves the goalposts your evidence is measured against.
• Policy environment: a new regulatory guidance document, or an HTA or payer policy change, that alters what evidence will be demanded of you.
• Independent research: a major publication that reframes the scientific question your asset is trying to answer.

Add two that a pharma-scale taxonomy underplays but every biotech lives with:

• Your own data: an unblinding, an interim, a safety signal. Your own readouts move the plan as much as anyone else's.
• Funding and partnering milestones: a raise, a partnering conversation, a board ask. Each puts the plan in front of an audience that will test it, so refresh before, not after.

The calendar becomes the backstop, not the driver. MAPS describes a hybrid "quarterly and ad hoc" model, and that is the right shape: a light-touch quarterly check that sweeps up whatever the triggers missed, with the event-triggered reviews doing the real work in between [1]. One trigger earns special mention. A regulatory interaction is among the highest-value events on the whole list, so if you are walking into an End-of-Phase-II meeting or a scientific advice procedure, refresh the plan on the way in, a moment that is genuinely make-or-break, not on whatever calendar date happens to fall nearby.

Fix three: a refresh is three questions, not a rebuild

The fear that kills good governance is that "review" means "redo." It does not. You built the plan once. The gap workshop that surfaced and ranked your evidence gaps, the audit that assessed where your evidence package stood, that heavy machinery runs once. A refresh touches it only at the edges, and it answers three questions:

1. Did a trigger fire that changes a priority evidence gap?
2. Does the ranked gap list still hold, or has something moved up, down, or off it?
3. Does any downstream decision, a study you were about to commission or a claim you were about to make, need re-flagging in light of the answer?

That is the whole exercise. If all three come back "no change," you close the review in an hour and lose nothing. If something moved, you have caught it while acting on it is still cheap.

Is this realistic for a stretched team? The living-guideline evidence says comfortably yes. Hewitt and colleagues reported in 2023 that weekly updating of guideline recommendations was feasible for a single standing group once the process was defined [9]. If a defined process makes weekly updating feasible for a national guideline, an event-triggered refresh two or three times a year is well within reach for a biotech running one asset. It asks for the same ruthless efficiency an IEP already demands when resources are tight, no more. Define the refresh narrowly and defend that definition: a short, bounded check against the trigger register, never the multi-week rebuild that produced the plan.

The strongest objection

Let me put the best case against my own argument, because it is a real one. Event-triggered review assumes somebody is continuously scanning the landscape. A one-asset biotech with no dedicated planning function does not obviously have that person. The calendar cadence, the objection runs, exists precisely because a fixed date is the only thing a stretched team will reliably act on. Event triggers sound rigorous, but with nobody watching they quietly degrade into no triggers at all, and a plan governed by no triggers is worse off than one reviewed, however bluntly, every quarter.

It is a fair challenge, and here is why the model survives it. The named owner running a register is less work than an annual re-derivation, not more: it swaps one big scheduled effort for a handful of small opportunistic ones. The triggers are not exotic monitoring tasks. They are things the team already watches go past, a competitor's readout, a guidance drop, its own unblinding. The register does not ask anyone to scan harder. It just turns what the team already clocks into something somebody has to act on. And the calendar is not abolished, only demoted, so the quarterly backstop still catches whatever slips through.

But the objection lands one honest hit, and I will not pretend otherwise. All of it depends on the owner genuinely having the time and the standing to do the job. Strip those away and event triggers really do collapse into nothing.

The one condition IEP governance depends on

Everything above is cheap to write down and hard to make real, for a single reason. A named owner with no protected time and no authority to convene a refresh is a title, not a control, and the plan will decay behind that person exactly as it did before the name existed. The only difference now is that the decay has someone's name attached to it.

So the thing to actually secure was never the governance document. It is that one person's mandate: the carved-out hours, and the standing to say "a trigger fired, we are reviewing the plan now" and have the organisation take it seriously. With that in place, the register and the cadence look after themselves. Building the plan around that mandate from the outset, rather than bolting governance on afterwards, is the kind of work we do at Inovia, InovaSight included, though the tooling is the easy part.

Miss the mandate, and you have simply written a better process for going stale.

References

1. Medical Affairs Professional Society (MAPS). "Strategic Integrated Evidence Generation Planning: Company and Non-Company Sponsored Research." Guidance Document, 11 February 2025. https://medicalaffairs.org/wp-content/uploads/2025/02/EVIDENCE-GENERATION-Standards-and-Guidance-2-11-2025.pdf
2. ZS Associates. Integrated evidence planning insights, including a proprietary leadership survey (April 2022; more than 30 leaders across 15 of the world's top 30 pharma companies). "Best practices for executing an integrated evidence plan" (https://www.zs.com/insights/best-practices-integrated-evidence-planning-and-generation); "Assessing your company's integrated evidence planning maturity" (https://www.zs.com/insights/assessing-your-companys-integrated-evidence-planning-maturity); "You've been asked to create an integrated evidence plan (IEP)—now what?" (https://www.zs.com/insights/you-ve-been-asked-create-integrated-evidence-plan-now-what).
3. Tendal B, Vogel JP, McDonald S, et al. (2021). "Weekly updates of national living evidence-based guidelines: methods for the Australian living guidelines for care of people with COVID-19." J Clin Epidemiol;131:11–21. PMID: 33188858. https://pubmed.ncbi.nlm.nih.gov/33188858/ — and Turner T, Elliott J, Tendal B, et al. (2022). "The Australian living guidelines for the clinical care of people with COVID-19: What worked, what didn't and why, a mixed methods process evaluation." PLoS ONE;17(1):e0261479. PMID: 34995312. https://pubmed.ncbi.nlm.nih.gov/34995312/ (The quoted sentence on cadence appears in Turner 2022.)
4. Shojania KG, et al. (2007). Survival analysis of how quickly systematic reviews go out of date. PMID: 17638714. https://pubmed.ncbi.nlm.nih.gov/17638714/
5. French SD, et al. (2005). Study of conclusion change on updating systematic reviews. PMID: 16225692. https://pubmed.ncbi.nlm.nih.gov/16225692/
6. Jaidee W, et al. (2010). Study of conclusion change on updating systematic reviews. PMID: 20644625. https://pubmed.ncbi.nlm.nih.gov/20644625/
7. Agarwal A, Rochwerg B, Lamontagne F, Siemieniuk RAC, Agoritsas T, et al. "A living WHO guideline on drugs for covid-19." BMJ 2020;370:m3379. doi:10.1136/bmj.m3379. PMID: 32887691. https://pubmed.ncbi.nlm.nih.gov/32887691/
8. Darley JM, Latané B. (1968). "Bystander intervention in emergencies: diffusion of responsibility." Journal of Personality and Social Psychology;8(4):377–383. doi:10.1037/h0025589. [Classic social-psychology study, cited strictly as an analogy; the headline helping rates of 85% / 62% / 31% are drawn from widely-corroborated secondary summaries of the 1968 experiment, not from a verbatim read of the primary paper.]
9. Hewitt J, McDonald S, Poole A, et al. (2023). "Weekly updating of guideline recommendations was feasible: the Australian National COVID-19 clinical evidence Taskforce." J Clin Epidemiol;155:131–136. PMID: 36813003. https://pubmed.ncbi.nlm.nih.gov/36813003/